On Tue, 27 Dec 2005, Dan Nicholson wrote:
bison, perl, vim - for me, these differ between my second and third
builds, but not for my third and fourth, nor for the first and second -
I guess this is the whole point of ICA and the build will be assumed
guilty unless an explanation can be produced. Needs investigation.
perl is a beast that I'll tackle another day. bison looks for bison
and flex, but will build without it:
My memory is getting hazy (been a long day), but I have a suspicion
that bison might also be affected by the missing gettext.
Unfortunately, I'm part way through rerunning my ppc ICA to get
dependable results (e.g. my in-place builds of perl linked -ldb because
I'd built blfs).
For vim I had oddities in configure (different results for pty ranges
and pty permissions), perhaps something to do with not mounting /dev in
the correct manner, so this is another reason for me to rerun. But at
the end of the day, my vim differences start to look like another eample
of this anonymous namespace stuff.
The big hassle is trying to make sense of differences - you'll know
that the diffs of binaries are at best hard to read (syntax highlighting
in view helps, but all the nulls ('@') don't. I tried using objdump -S,
but I don't feel confident in saying that if two code disassemblies
match, the files are close enough - the remainder (text section) can be
very different. Using objdump -S -D didn't seem practical - it's very
slow, and the text section is shown as a hexdump of 16 (?) bytes per
line - diff can't really cope with trying to show what has changed
within it in an easy-to-read way, and anyway objdump -S -D is *very*
slow to execute. I suppose saying that the code matches (or not) might
be an advance on "this usually differs".
I'm gonna try to get flex linked
into bison and see if that cleans up the differences. Do you think I
should move flex up in Ch. 6 or add bison and flex to Ch. 5?
We all know how much everybody groans at yet another package in chapter
5, so if it's fixed by reordering chapter 6, that would be more
acceptable.
Ken
--
das eine Mal als Tragödie, das andere Mal als Farce
--
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
